Weekly AI search visibility brief for your marketing team

Every Monday at 8am, build a competitive AI search visibility brief for my marketing team, post it to Slack, and log the numbers to a Google Sheets tracker.

Start with Profound's Query Visibility V2 report. Pull share of voice, visibility score, and average position for my brand and every competitor I track, covering the last 7 days, with the prior 7 days set as the comparison window so week over week deltas come back from the report itself rather than needing stored history. Break the results down by model so ChatGPT, Perplexity, and Google AI Overviews each show up separately instead of being blended into one number. Profound's v2 report date ranges are inclusive on both ends, so set the start and end dates accordingly.

Then use Profound's Query Citations V2 report for the same 7 day window to pull the top citing domains and pages behind those answers, including citation count, citation share, and rank. This is the evidence layer that explains why visibility moved, so keep it tied to the brands and models from the visibility pull.

Have the agent write a short narrative brief, not a metric dump. It should name which brands gained and which lost share of voice, and call out only moves larger than 2 points so the brief stays signal heavy. For each move worth mentioning, explain which sources look like they drove it, using the citation data to point at the specific domains or pages that gained or lost ground. Flag any model where the picture differs sharply from the others, since that is usually where the actionable gap is. Close with two or three concrete recommended actions the marketing team can take this week.

Post the finished brief into my Slack channel using Send a Message. Format it for Slack readability with short sections and bold headers, aimed at something a marketer will actually read in the channel rather than a wall of numbers.

Also append one row per brand per week to my Google Sheets tracker using Append Values. Each row should carry the week ending date, brand name, model (or a blended all models row), share of voice, visibility score, average position, and the change in share of voice versus the prior week. This gives the team its own visibility trendline over time instead of relying on whatever window the dashboard happens to show.

Two practical notes. Profound API access is rate limited to 600 requests per hour per key, so keep the number of report calls modest and page through results rather than firing one call per brand per model where that can be avoided. If Profound returns no data for the window, post a short note to Slack saying so instead of publishing an empty brief, and skip the spreadsheet append for that week.